PSU diode replacement.. fact or fiction?

ok.. yes i've read this topic somewhere here or at the old place, but my searches turned up 821 topics.. the first 30 didn't have anything to do with what I'm wondering here..

What's with the audiofools changing PS diodes from standard to some kind of zener? I read that it's supposed to cut down on noise and make the thing more efficient but does it really? I've got a switching PSU.. now do you really think rectifying diodes on the front end would help at all? I think i need more :guinness: before i start to believe that.. So I suppose we narrow it down to linear PSUs, now would they help them much?

any hard evidence or professional opinions supporting this? Clips of before and afters? i don't think "i swear i can hear a difference" really proves anything to me... :green:

Um I doubt they switch rectifier diodes to zeners...sounds like abad idea. But they do switch them out for Hexfred soft recovery and low leakage diodes. I suppose they could be a little better than the standard diodes. They're cheap, so it could be worth an experiment or two....

As far as using them in switching supplies....I don't know, that;s too complicated to even hazzard a guess at what different diodes might do. Personally I stay away from switchers wherever possible...they seem to leak 100kHz and up into everything they feed. And it's almost impossible to tame that.

Cheers.

Kris

Zeners will fry.
They means either FREDS, like Kris says, or sometimes Schottky diodes, which do not come in a high enough voltage for a vacuum tube supply, but do come in voltage ratings high enough for a transistor supply.
If you have all your other noise sources at their minimum but still want more reduction, give them a try. But they are probably down the list at 3 or 4 as far as noise sources are concearned. Pwr supply hum, heater hum, tube and transistor noise, grounding problems, are probably going to be many db above rectifiers as far as overall noise goes.
Thats why I don't bother with them. They are right on the border of audiophoolery in my opinion, but here are some numbers anyway:

1N3890 12A 100V 200NS (NANO SECONDS)
1N4933-1N4937 FAST RECOVERY 1A 50V 200NS TO 1A 600V 200 NS
1N5817 SCHOTTKY 1A 20V
1N5822 SCHOTTKY 3A 40V

CJ

hmmm.. sorry, I did mean hexfreds and the like but my day has been so bad that all that came out was ZENER! that's all i dealt with today was a design using zeners that is utterly moronic and worthless, yet the boss likes it and is forcing me to MAKE it work... :shock:

thanks for the info!

I have built some bridge rectifiers from Hexfred diodes for various small consoles and power amps. I have the very subjective impression that things were a bit cleaner and quieter. It was a very subtle difference, though, and it could certainly have been wishful thinking.

:thumb: :thumb:

I guess you could series the zeners if you do not want them to conduct in reverse, as they do when used as voltage regulators.
Your power rating for zeners is typically one watt or 1/2 watt, but thats when conductiing current at their zener voltage, which is going to chew up zener volts times amps, so say a 12 volt 1 watt zener is used, you can only conduct 1/12 amps in reverse. But if they do not conduct, you only have about 0.6 volts times current, so the same zener should give you 1/0.6 equals 1.66 amps, derate 50 percent for saftey, 800 ma.

Do these "improvements" have to do with switching speed reducing the slew rate, or is there some other mojo at work in those audiophool diodes?

Joel

well the design i'm working on is actually for a digital signal being modulated on an ac signal.. this circuit strips the digital off and feeds it to an MCU.. I asked the question earlier for no real reason other than curiosity!

thanks guys!

One thing he may be thinking of is: He said he was using a switcher.
You can put a bridge rectifier and a filter section to make D.C. and then run the switcher on of D.C. for better ripple and less EMI artifacts generated from the switcher. Most switchers (unless they are designed for audio) have a crappy front end and absolutly the minimum amount of capacitance. It'll have just enough to get the job done.

RonL

PS.
Don't put the switcher in the same box with the audio stuff!

Rectifier diodes in PSU can produce RF interference when recovering from reverse bias. This has some influence on performance and also FCC may have something to say. That's why you see small caps ca 10nF parallelled with the diodes. The recovery caused RF interference can be often seen in tube circuits where a couple of amperes of heater supply is rectified. Sounds like crackling noise.
Hexfreds are better in that respect.

I use BYV27-200 for my psu's! (2A) There's a BYV27-600 (1,6A) maybe useful for tube builders?
They all goes under the name "Ultra fast epitaxial rectifiers"
If you use this type of diods, don't put a capacitor over them! then you'll end up with more recovery spikes!
Farnell and Newark carry them!

www.fairchildsemi.com/ds/UF/UF4002.pdf

http://www.vishay.com/diodes/list/product-88755/

[quote author="Svart"]hmmm.. sorry, I did mean hexfreds and the like but my day has been so bad that all that came out was ZENER! [/quote]

Zener diode is highly-dopped diode and have then high conductivity and
low series resistancy if used as conventional diode.
It can be useful in some like clone FAIRCHILD compressor,
but in the power supply? it may be only source of buzz.

xvlk

[quote author="Tekay"]I use BYV27-200 for my psu's! (2A) There's a BYV27-600 (1,6A) maybe useful for tube builders?
They all goes under the name "Ultra fast epitaxial rectifiers"
If you use this type of diods, don't put a capacitor over them! then you'll end up with more recovery spikes!
[/quote]

You shouldn't just put a cap across a diode anyways. You should use a proper snubber network instead. Parrallel the diode with a resistor and cap in series, that is what they do in large power diodes to help with turn off spikes. There is a really good paper on snubber networks, but I can't remember where I found it. I'll look around for it if anyone is interested.

i'm interested. i use snubbers a lot in my job.

"Calculating optimum snubbers" by Jim Hagerman might be it.

I have it as a pdf I can squirt you, heaven know where I got it from, maybe google the title.

[quote author="Jonathan Hayward"]"Calculating optimum snubbers" by Jim Hagerman might be it.

I have it as a pdf I can squirt you, heaven know where I got it from, maybe google the title.[/quote]

Yep, that was what I was thinking of.

http://peufeu.free.fr/audio/articles/Snubber%20Networks%20in%20Power%20Supplies.pdf

I think that link should do.

Otherwise click here- snubberpaper